by T. R. Reid
An affable, naturally gregarious person, Noyce seemed to gravitate to a leadership position in most things he undertook. In the early 1980s, he got interested in madrigal singing. He studied, practiced, and joined a chorus; soon enough, he became its conductor. A friend once took Noyce to an Audubon Society meeting; soon enough, he was supervising a global effort to find a new habitat for a dwindling species of Newfoundland puffin. He was one member of a large group of Silicon Valley executives who decided a few years back that the microelectronics business needed a professional organization; soon enough, Noyce was chairman of the Semiconductor Industry Association.
A wiry, athletic-looking man with angular features and curly hair that went silvery gray in his sixties, Noyce carried himself with the rakish self-assurance of a racing driver or jet pilot (despite a clear memory of that first disaster, Noyce was always fascinated with airplanes and flew a rebuilt Republic Seabee, a World War II seaplane). A fashionable dresser, he wore silver-rimmed aviator glasses, a large gold ring, and a platinum watch that had both old-style hands and new-style digital readout. If Jack Kilby seems as loose as a spare piece of string, Noyce was a coiled spring of energy and enthusiasm. When the Harvard Business Review asked him to describe the employees of Intel, the massively successful microchip company he cofounded, Noyce responded with an answer that might serve as a self-portrait. “They’re high achievers,” he said. “High achievers love to be measured, when you really come down to it, because otherwise they can’t prove to themselves that they’re achieving.” Among such people, Noyce added, “I don’t think you could call it a relaxed atmosphere. A confident environment, but not a relaxed one.” A confident, but rarely relaxed, high achiever—that description fit Bob Noyce perfectly. One of Noyce’s acolytes at Intel, Andrew Grove, who took over as chairman of the firm in the 1990s, offered a somewhat more chilling description of the high-tension corporate culture: “Only the paranoid survive.”
One reason it was difficult to relax at Intel when Noyce was running the place was that the firm deliberately kept itself on the leading edge of technology, always looking for the new idea. Actually, that’s the way Intel still operates, more than a dozen years since the patriarch left. That’s the way Robert Noyce liked it; he defined himself as a “technologist,” and defined that term as “the kind of person who is comfortable with risk.” And there lies the key difference between a technologist and a businessman, Noyce said: “No businessman would have developed the telephone. It’s got to be a maverick—some guy who’s been working with the deaf and gets the crazy idea that you could actually send the human voice over a wire. . . . A businessman would have been out taking a market survey, and since it was a nonexistent product, he would have proven conclusively that the market for a telephone was zero.”
The same description—“comfortable with risk”—applied to Noyce’s approach to problem solving. He attacked engineering problems with absolute confidence that there was a perfectly good solution somewhere, and that the human mind could find it. He was to be a prolific, impulsive producer of solutions himself, but he also knew that “a lot of my great ideas turn out to be ridiculous when you look closely.” For that reason, Noyce was never a solitary inventor like Jack Kilby. He needed to work with others, to talk things over, to launch his ideas and see if they would fly. “You explain a lot of things to yourself,” he said, “by trying to explain them to someone else. And then either you can see for yourself that the idea won’t work, or the other person can spot the problem and help you find a better way.”
Early in his career Noyce came into the ken of another famous technical optimist, William B. Shockley, and it is clear that he absorbed Shockley’s first rule for solving problems: “Try simplest cases.” Asked to describe the right way to find a solution, Noyce put it this way: “All of our progress, in any technical field—well, there are a few leaps of insight—but beyond that it has been decomposing it to its simplest elements, to understandable elements, and building it back up from its simplest element. Shockley had this wonderful ability to make the right simplifying assumption, to get the math out of the way until you had a basic visual image of what was happening. . . . Well, you see, you have to get back to the basic, the simplest picture before you can understand a problem well enough to solve it.”
Like Kilby, Noyce tended to be instantly suspicious of any solution that seemed too obvious. “It’s absolutely true, whether it’s a technical question or anything else, that a lot of people are going to try to approach a problem the same way everybody else has,” Noyce said. “You can just let yourself get in a rut. And if you don’t get out of the rut, you’re not going to solve the problem.” The person who can jump over that rut—that person, the person who can be a successful inventor—is the one who is able to come up with the “leap of insight.” “The successful solution comes about because somebody was able to fire up his imagination and try something new . . . ,” Noyce added. “If you want to achieve something worthwhile, you have to jump to the new idea.”
Noyce spent his entire boyhood pursuing new ideas, new phenomena, new gadgets. He was born in 1927 in the tiny town of Denmark, Iowa, the son of a Congregationalist minister. The family moved from one small rural town to the next when Bob was small; when his father took over the parish in Grinnell, a quiet, attractive town fifty miles east of Des Moines, the family settled there for good. Although the elder Noyces had no particular penchant for technology, they encouraged their sons and listened with interest each night as Bob and his three brothers explained their latest experiments. To pay for his airplane kits, radio parts, chemicals, etc., Bob worked as a baby-sitter and mower of lawns. One of his chief customers was Professor Grant Gale, chairman of the physics department at Grinnell College. Under Gale’s tutelage, Noyce fell in love with math and physics. He studied the college texts while in high school, and when he enrolled at Grinnell in 1945 he knew from the first that physics and math would be his major interests. Not his only interests, of course—that was not the way Bob Noyce lived. He was the star diver on Grinnell’s swimming team, he sang in choral groups, played oboe in a band, and had a continuing role in a radio soap opera.
Noyce’s life story was basically a narrative of one outstanding success after another, but at college he experienced an unforgettable setback. One night in 1948 a group of Grinnell boys decided—it evidently seemed logical at the time—that what their dormitory in the middle of the Iowa plains really needed was a genuine Hawaiian luau, complete with roast whole suckling pig. Tasks were divvied up: Bob and another athletic sort were assigned the job of acquiring the pig—a mission they accomplished by swiping a suckling out of the sty at a nearby farm. The Iowan luau was a great success. But the next morning, when the repentant Noyce confessed his crime, nobody was laughing. In Iowa, theft of livestock is not a humorous matter. Thanks to his father’s stature and Grant Gale’s intervention, the pork thief was spared a criminal sentence and was kicked out of college for only one semester. He spent the time working at an insurance company. When he reentered Grinnell midway through his senior year he was still able to graduate with his class, earning top grades and making Phi Beta Kappa.
For the most part, the electronics Noyce learned in his physics classes at Grinnell was standard vacuum tube stuff—the Fleming rectifier, the De Forest amplifier, and various improvements to those basic devices. One day, however, Professor Gale astounded the class with news of something totally different. Gale had been a classmate of John Bardeen in the engineering school at the University of Wisconsin, and thus he was able to obtain one of the first transistors and demonstrate it to his students. It was not a lecture the student was to forget. “It hit me like the atom bomb,” Noyce recalled forty years later. “It was simply astonishing. Just the whole concept, that you could get amplification without a vacuum. It was one of those ideas that just jolts you out of the rut, gets you thinking in a different way.”
Solid-state technology was a whole new world, a whole new universe, f
or a student of physics. Noyce was bright enough to realize that this had to be the future; there would be no going back to the vacuum tube. As a student of physics at precisely the right moment, Noyce decided to enter this astonishing new world. He applied to MIT and was admitted to a graduate course in physical electronics—and was accepted, of course. He took his Ph.D. there in 1953, sifted through a long list of job offers, and went to work in Philadelphia for Philco, which was just embarking on a large-scale effort to develop and produce improved transistors. The job gave Noyce a chance to practice serious science. He turned out a series of monographs and papers on semiconductor devices. In one paper, Noyce reported on the effects of low-energy gas discharges on a platinum-germanium diode; in another, he demonstrated that the “dc transistor current amplification factor” could be determined using this formula:
Late in 1955, Noyce gave a paper before the American Physical Society on “base widening punch-through.” That paper caught the eye of the nation’s preeminent semiconductor specialist.
On a January day in 1956, Noyce received a telephone call from William B. Shockley. Shockley explained that he was leaving Bell Labs and moving to California to start a new company that would develop high-performance transistors. Would Noyce be interested in interviewing for a job? “It was like picking up the phone and talking to God,” Noyce recalled later. “He was absolutely the most important person in semiconductor electronics. Getting that job meant you would definitely be playing in the big leagues.” Noyce took a cross-country train to Palo Alto. With characteristic Noycean confidence, he spent the morning of his arrival renting a house near Shockley’s lab; that done, he went to the interview to see if he could land the job. He got it, settled his family into the new home, and set to work with Shockley on the development of a high-performance double-diffusion transistor.
The transistor, as Shockley conceived it, is a semiconductor sandwich in which a thin region of P-type silicon (that is, silicon that has been doped with impurities so that it contains extra positive charges) is sandwiched between two regions of N-type silicon (silicon doped with excess negative charges). (There are also less common transistors that employ the opposite structure, with N-type material in the center and P-type on either end.) To get the best performance characteristics from the device, the separate regions of the silicon chip should be clearly defined, and the P-N junction, the point where the different regions meet, should be a sharp, sudden transition from P-type to N-type material. Through the early 1950s transistor makers tried dozens of different techniques to achieve the precise doping and the sharply defined junctions required for reliable transistor action. The process that eventually proved best—the process still used today in semiconductor manufacture—was a Bell Labs discovery called diffusion. Semiconductor diffusion works like a barbecue pit where hickory smoke seeps into the meat and imparts a distinctive flavor. In the diffusion process, a bar of silicon is cooked in a furnace at high heat, and then a gas containing the appropriate doping impurities—boron, for example, or arsenic—is pumped into the furnace. At temperatures of 1000 degrees centigrade or so, some of the impurity atoms in the gas seep into the silicon bar and dope it, either with excess electrons or with positively charged holes. In the same way that a barbecue chef knows just how long to cook the ribs to get the right taste of hickory, solid-state physicists gradually determined the proper time and temperature needed to put precise amounts of impurities at precise points on the silicon block. Thus diffusion provided the first effective means of producing a semiconductor bar with sharply defined regions of N- and P-type silicon. The process used to make transistors was called double diffusion.
A round wafer of silicon, about the size of a 45 rpm record in diameter and about the thickness of five pages of this book, would be placed in the furnace. Two different kinds of impurities would be diffused onto the wafer in separate steps, leaving a three-layer cake—N-type on the bottom, P-type in the middle, N-type on top. The wafer could then be cut, like a cake, into dozens or scores of tiny three-layer pieces—each one an N-P-N transistor.
Double diffusion made possible, for the first time, the mass production of precise, high-performance transistors. The technique promised to be highly profitable for any organization that could master its technical intricacies. Shockley therefore quit Bell Labs and, with financial backing from Arnold Beckman, president of a prestigious maker of scientific instruments, started a company to produce double-diffusion transistors. The inventor recruited the best young minds he could find, including Noyce; Gordon Moore, a physical chemist from Johns Hopkins; and Jean Hoerni, a Swiss-born physicist whose strength was in theory. Already thinking about human intelligence, Shockley made each of his recruits take a battery of psychological tests. The results described Noyce as an introvert, a conclusion so ludicrous that it should have told Shockley something about the value of such tests. Early in 1956, Shockley Semiconductor Laboratories opened for business in the sunny valley south of Palo Alto. It was the first electronics firm in what was to become Silicon Valley.
In Robert Noyce’s office there hung a black-and-white photo that showed a jovial crew of young scientists offering a champagne toast to the smiling William Shockley. The picture was taken on November 1, 1956, a few hours after the news of Shockley’s Nobel Prize had reached Palo Alto. By the time that happy picture was taken, however, Shockley Semiconductor Laboratories was a chaotic and thoroughly unhappy place. For all his technical expertise, Shockley had proven to be an inexpert manager. He was continually shifting his researchers from one job to another; he couldn’t seem to make up his mind what, if anything, the company was trying to produce.
“There was a group that worked for Shockley that was pretty unhappy,” Noyce recalled many years later. “And that group went to Beckman and said, hey, this isn’t working. . . . About that time, Shockley got his Nobel Prize. And Beckman was sort of between the devil and the deep blue sea. He couldn’t fire Shockley, who had just gotten this great international honor, but he had to change the management or else everyone else would leave.” In the end, Beckman stuck with Shockley—and paid a huge price.
Confused and frustrated, eight of the young scientists, including Noyce, Moore, and Hoerni, decided to look for another place to work. That first group—Shockley called them “the traitorous eight”—turned out to be pioneers, for they established a pattern that has been followed time and again in Silicon Valley ever since. They decided to offer themselves as a team to whichever employer made the best offer. Word of this unusual proposal reached an investment banker in New York, who offered a counterproposal: Instead of working for somebody else, the eight scientists should start their own firm. The banker knew of an investor who would provide the backing—the Fairchild Camera and Instrument Corporation, which had been looking hard for an entrée to the transistor business. A deal was struck. Each of the eight young scientists put up $500 in earnest money, the corporate angel put up all the rest, and early in 1957 the Fairchild Semiconductor Corporation opened for business, a mile or so down the road from Shockley’s operation.
Noyce recalled that the group had some slight qualms about running their own business, but these doubts were easily overcome by “the realization, for the first time, that you had a chance at making more money than you ever dreamed of.” The dream, as it happened, came true. Even by high-tech standards, that $500 turned out to be a spectacular investment. In 1968 the founders sold their share of Fairchild Semiconductor back to the parent company; Noyce’s proceeds—the return on his initial $500 investment—came to $250,000. Noyce and his friend Gordon Moore had by then found another financial backer and started a new firm, Intel Corporation (the name is a play on both Intelligence and Integrated Electronics). Intel started out making chips for computer memories, a business that took off like a rocket. Intel’s shares were traded publicly for the first time in 1971—on the same day, coincidentally, that Playboy Enterprises went public. On that first day, stock in the two firms was about equally priced;
a year later, Intel’s shares were worth more than twice as much as Playboy’s. “Wall Street has spoken,” an investment analyst observed. “It’s memories over mammaries.” Today, Intel is a multibillion-dollar company, and anybody who held on to the founding group’s stake in the company is a billionaire several times over.
The men who started Fairchild Semiconductor in 1957 were determined to make the double-diffusion transistor, but they were also on the outlook for any other product that could turn a profit. Noyce, as director of research and development, was the man responsible for spotting important technical problems that Fairchild might profitably solve. And it didn’t take a genius, at this point in technological history, to identify the most important problem by far facing the electronics industry: the tyranny of numbers. Off and on, all through 1957 and 1958, Noyce thought about the interconnections problem. In retrospect, he said later, the monolithic idea should have come to him much earlier. “Here we were in a factory that was making all these transistors in a perfect array on a single wafer,” Noyce said, “and then we cut them apart into tiny pieces and had to hire thousands of women with tweezers to pick them up and try to wire them together. It just seemed so stupid. It’s expensive, it’s unreliable, it clearly limits the complexity of the circuits you can build. It was an acute problem. The answer, of course, was, don’t cut them apart in the first place. But nobody realized that then.” Instead, Noyce was stuck in a rut. He worked on standard ideas for making circuit components in smaller sizes and uniform shapes. He came up with nothing worthwhile.